When undergoing treatment for breast cancer, many women experience severe side effects that can persist for years, and toxicities often result in reduced dose and efficacy. Little is understood regarding factors that may predict drug toxicities. Pharmacogenetics has been used to determine susceptibility to treatment-related toxicities, using a candidate gene approach, and some important findings have been made in relation to metabolism of thiopurines and irinotecan. There has been less progress in assessment of genetic variants that predispose to toxicities resulting from a multi-drug regimen of cyclophosphamide (C), anthracyclines (A), and taxanes (T), commonly used to treat breast cancer. We propose to conduct a Genome-Wide Scan (GWAS) in an ongoing clinical trial for breast cancer, and examine genetic variants in relation to Grade 3 and 4 toxicities. Using data and samples (n=2000) from a large therapeutic trial (S0221), we will perform a GWAS to identify SNPs significantly associated with grades 3 and 4 hematological and gastrointestinal toxicities during the AC segment, and neurological toxicities during the T segment. Results from S0221 GWAS will be validated in CALGB 40101, a trial of 2200 women with similar entry criteria receiving the same chemotherapy agents. Data from the two trials will also be pooled for meta-analysis, enhancing statistical power. By using a GWAS approach, it is likely that important pathways not previously considered will be revealed as important in susceptibility to treatment-related toxicities, identifying those at greatest risk for alternate drugs or dose reduction, and opening new areas of research for prevention of often life-threatening toxicities among patients receiving chemotherapy for breast cancer.